82 
EIGHTH REPORT. 
tions given are as near as can be learned those of the gages above sea 
level. At Weather Bureau stations therefore the ground surface is one 
or two hundred feet lower. Such stations are marked WB in the margin 
of the list. 
Fig. 3 has been prepared to show at a glance the rainfall at each sta¬ 
tion made use of, thus reduced to uniform epoch, as well as the length 
of the series of observations by the small number written alongside. 
This and the table of data that follows give the reader all the facts 
used in preparing our chart of rainfall. How the isohyetals or lines 
of equal rainfall have been drawn will be best made out by inspection 
of Fig. 3. In the main the figures have been closely followed with some 
slight modification for probable error of the series of different lengths. 
This matter of the probable error of the mean of a series of observa¬ 
tions of precipitation of any given length has not yet been sufficiently 
studied. There are some considerations cited in Ilann’s Meteorologie 
at p. 324, from which it appears that a rainfall series of ten years 
may be liable to an error of about five inches in our neighborhood, 
one of twenty-five years about two inches. The accordance of neigh¬ 
boring stations may greatly reduce their liability to error. 
